[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [ARSCLIST] Advice for filtering surface noise
On 01/10/04, Goran Finnberg wrote:
> Don Cox:
>> How big, exactly, is "an order of magnitude" ?
> 10 to 30 dB depending on lateral cutting velocity, curvature error
> depending on frequency and diameter of the cut together with if the
> replay stylus is conical versus elliptical.
I thought you might be using the term vaguely. An order of magnitude is
2x if you are using binary, or 10x if you are using decimal.
Now, let's concentrate on the surface defects of a mono record. The
groove walls are at roughly 45 degrees.
Draw a line at 45 degrees. Now, the mono signal can be represented by a
horizontal vector, while the surface imperfections can be represented by
a vector normal to the 45 degree line (or to the groove wall).
That noise vector can be resolved into two equal components, horizontal
and vertical. It follows from Pythagoras that the original vector is
root 2 times each of its components. How you can get 10dB out of that
NB I am not talking about pinch distortion or subsonic rumble.